Descriptive notation
Descriptive notation was the most common form of notation used to record a chess game until late in the 20th century. As such, while it has been supplanted in modern use by algebraic notation a knowledge of descriptive notation remains important as a large wealth of classical chess literature uses it and much of it has not been republished in the newer format. In addition, it can sometimes be useful to reference squares in descriptive as noted below. In descriptive notation, pieces are referred to mainly by the same letters as they are in the newer algebraic notation. The main exceptions are the knight and pawn; the former is often referred to as Kt (although the modern N is also valid) and the latter, which nominally lacks an abbreviation in algebraic is explicitly referred to as P in descriptive. The eight files are named based on the lateral side (Queen's or King's) of the board they belong to and the name of the piece that starts the game on that file. For instance, the file on which the queen's side rooks begin the game on is called the Queen's Rook's file, abbreviated QR; the file that the king's side knights begin the game on would be the King's Knight's or KKt (alternatively KN) file. The files the queen and king themselves start the game on are of course simply the Queen and King files with their single letter abbreviations. The eight ranks are numbered 1-8. For White, they are the same as in algebraic notation, however, descriptive notation is perspective based, so for Black they are numbered beginning at Black's back row, in other words White's 8 is Black's 1 and vice versa, and similarly 2 and 7, 3 and 6, 4 and 5 are inverted for Black as compared to White. A square's full name, thus, is the combination of its file name (or abbreviation) and its rank number from the player's perspective. Thus, for example, the square that is familiar to players as d3 in algebraic notation, would be on the Queen file in descriptive. For White it would be Q3; for Black we invert the perspective and thus from his perspective it is Q6. The square f2 is on the King's Bishop's file, and thus is KB2 for White, but of course KB7 for Black. And so on. (Thus is perhaps the main usefulness of descriptive notation in modern times. Most attack tactics tend to work the same from either end of the board, and thus it is sometimes efficient to refer to a square from a generic perspective. For instance, while classical literature would sometimes refer to a bishop sacrifice at KR7, modern notation would require that either a particular player's perspective be assumed or the somewhat clunky "h7 or h2" be stated instead.) Note that in practice the Queen's or King's prefix is not normally given to a square when describing a move unless it is necessary for disambiguation. For instance, the aforementioned f2 would probably usually simply be called B2 or B7, unless the circumstances of a given move made it necessary to specify the King's side. While in descriptive notation, destination squares are still given for non-capturing moves, capturing moves are instead described as the pieces involved. For instance, White captures a pawn on e5 with a knight on f3 - in algebraic, this is written Nxe5; in descriptive it is instead written NxP (not NxK5). This is of course assuming that only one White knight can capture a pawn, and that that is the only pawn that the knight can capture (otherwise, we use disambiguations, as noted below). En passant captures are noted as "PxP e.p." Disambiguation between pieces in descriptive notation (whether for the moving piece, or for the victim of a capture) is provided by the following means: * Knights and Rooks are referred to either by their lateral side (i.e., in the early game where it is still possible to recall which Knight or Rook was the Queen's or King's respectively), or can be referred to by their rank or file (sometimes, the whole square name is given) if that is no longer reasonably feasible. For instance, in the diagram referred to as Example 1, consider the algebraic move Ncd1. Disambiguation between the white knights is needed. If we can still be reasonably sure that the knight on c3 was the Queen's Knight, then we can (and most likely will) simply write the move as QN-Q1. If we are not sure, then we simply specify where the Knight came from. Here either N(QB)-Q1, N(B3)-Q1, N(3)-Q1, or N(QB3)-Q1 is acceptable (most commonly, N(B3)-Q1 would be chosen - there is no White knight on KB3, nor would it be possible for such a knight to move to Q1 in any event so saying QB3 is redundant, while N(QB)-Q1 is frowned upon since it's customary to refer to a file and rank, not just a file. However, N(3)-Q1 is almost as common, as giving the rank alone is not considered as inelegant as giving just the file). * Pawns are specified by the name of the file they are on, normally omitting the Queen's or King's side designator unless necessary. Again to take the diagram Example 1 as an example, say White plays what we would call Nxg4 in descriptive notation. We can't say just NxP - there would be no way to tell then whether we meant the KP or the KNP (not to mention, since both White knights in this diagram can capture pawns, we would leave the reader unsure as to which knight, too). We have to specify which pawn so we would write NxKNP or NxNP - since there is only one Knight's Pawn that can be captured by a knight, we say simply NxNP. (If Black's pawn at b6 were at b5 in the diagram, then we would really have to say NxKNP.) Or consider the move Ncxe4 in this diagram, here we have to specify both the knight and the pawn, for both White knights can take the e4 pawn and both can also take more than one pawn. In addition, both of the pawns the knight on c3 can take are on the King's file, so N(B3)xKP won't do, either. Here we would simply specify the square for both affected pieces: N(B3)xP(K4) or N(B3)xP(4). Note that since pawns only move along their files unless capturing, a simple pawn move is written simply as P- and the square name, e.g., if Black were to play e3, the move is simply P-K6, and there is no way that this could cause ambiguity issues. Of course b3 or g3 would have to be distinguished as P-QN6 or P-KN6 since P-N6 would leave the reader in the dark as to what side of the board this is happening on, but it is never necessary in a simple pawn move to refer to the pawn as anything but simply P. Pawn captures of course sometimes need disambiguation, since two pawns could capture the same piece. For instance, if Black plays exf3 here, one would simply say KPxR. Note that gxf3 would be written differently because NPxR could refer to either gxf3 or bxa3, the most usual way to write it would be NPxR(B6). (Although NPxR(B) is actually sufficient here, it is customary to give the rank as well in such cases.) Of course, bxc3 would simply be PxN, as there is only one Black pawn that can capture a knight, and the fact that it could have taken the rook instead is irrelevant. * Bishops, because of their monochromic nature are almost always referred to throughout the game as QB and KB. For example, the moves Bxb6 and Bxg4 in the given diagram would be written simply QBxP and KBxP respectively. However, it is customary here, as with Pawns, to disambiguate the square rather than the piece during a non-capturing move; therefore, Bc5 and Bf5 here would be B-QB5 and B-KB5 respectively, rather than QB-B5 and KB-B5 (although the latter are technically also valid). The only real exception to this is the rare case in which an underpromoted bishop is present, in which case we simply specify squares as noted. Similarly, in cases where there is more than one queen of a color on the board due to promotion, if disambiguation is needed we simply refer to the square that the referenced queen was on. Check can be referred to by + as in algebraic notation (some books such as Bobby Fischer's My 60 Memorable Games do so), but is more commonly abbreviated as "ch." Checkmate is normally simply "mate." Some publications distinguish discovered and double checks (via "dis ch" or "dbl ch") but this is not strictly necessary. Finally, there are some older publications which use alternate descriptions for squares on a player's 1st rank. Sometimes "1" is called "sq" (e.g. "Q sq." which is equivalent to Q1) and sometimes the rank is omitted entirely. (For instance, B-K would be equivalent to B-K1.). This however is quite rare and is primarily seen only in particularly old literature.